The aim of this work is to constrain the mean SFRs as a function of AGN bolometric luminosity, reaching very high luminosities (LAGN∼ 1048erg s−1; see Fig. 4.1), as well as
investigating dependencies of the mean SFRs on the presence of a radio-luminous AGN. Far-IR (FIR) photometry is one of the best tracers of star formation, as it traces the peak of the dust-reprocessed emission from star-forming regions (e.g. Kennicutt 1998; Calzetti et al. 2010; Dom´ınguez S´anchez et al. 2014b). We use FIR data from the Her- schel-ATLAS observational program (H-ATLAS; Eales et al. 2010a; section 4.2.2) that covered the fields of GAMA09, GAMA12, and GAMMA15 in its Phase 1, and the north and south galactic poles (NGP, and SGP respectively) in its Phase 2 observations. The Herschel-ATLAS fields benefit from multi-wavelength coverage, with excellent optical (SDSS; section 4.2.1), MIR and FIR photometry (WISE and Herschel; section 4.2.2), and radio observations (FIRST; section 4.2.3). We use the available data to draw a sample of optically selected QSOs from the SDSS survey, determine a radio-luminous sub-sample of QSOs using the FIRST survey, and define their SFRs using the WISE and Herschel observations. As we only study the fields that have overlap with the SDSS survey area, we exclude the SGP field.
4.2.1
Optical/SDSS QSOs
To define our QSO sample we use the publicly available SDSS data release 7 (DR7) QSO catalogue as presented in Shen et al. (2011) (see also Schneider et al. 2010 for original selection of QSOs). To provide a measurement of the power of the QSOs we use the AGN bolometric luminosity as given in this catalogue. This luminosity has been derived from
4.2. Sample & Data used 97
Figure 4.1: AGN bolometric luminosity (LAGN) versus redshift (z) for the full QSO
sample from SDSS DR7 covered by H-ATLAS in the NGP, GAMA9, GAMA12, and GAMA15 fields. The vertical dashed lines indicate the redshift ranges taken in our anal- ysis, and the horizontal dashed line shows the LAGNcut that defines the sample (see sec-
tion 4.2.1). In red we highlight the radio detected sources from the FIRST radio catalogue (see section 4.2.3). Within the redshift range of interest (z =0.2–2.5) there are a total of 3026 optically selected QSOs.
Figure 4.2: Radio luminosity from the FIRST survey (L1.4GHz) versus redshift (z), for the
full QSO sample from SDSS DR7 covered by H-ATLAS in the NGP, GAMA9, GAMA12, and GAMA15 fields that is radio detected. The vertical dashed lines indicate the redshift ranges taken in our analysis, and the horizontal dashed lines show the L1.4GHzlimits used
to define sources as radio-luminous. A total of 258 are classified as radio-luminous within the redshift range of interest (z =0.2–2.5; see section 4.2.3).
4.2. Sample & Data used 99 L5100A, L3000A, and L1350A, for sources at redshifts of z <0.7, 0.7≤ z <1.9, and z ≥1.9
respectively, using the spectral fits and bolometric corrections from the composite SED in Richards et al. (2006) (BC5100A= 9.26, BC3000A= 5.15 and BC1350A= 3.81; see Shen
et al. 2011). We constrain the sample of QSOs within the regions covered by H-ATLAS, removing sources that are close to the image boundaries. All the QSOs of our sample have a bolometric luminosities of LAGN& 1045erg s−1(see Fig. 1).
We also make use of the virial BH masses (MBH) estimates from Shen et al. (2011),
from which we estimate the stellar masses (see section 4.4.2 and Eg. 4.4.4). The MBH
have been calculated using the FWHM of Hβ, MgII, and CIVlines (see section 3 of Shen
et al. 2011). Specifically, the MBH is estimated from Hβ for sources with redshifts of
z<0.7, from MgIIfor sources with 0.7< z ≤1.9, and from CIVfor sources with z >1.9.
This study looks at sources with redshifts z = 0.2–2.5, and includes a total of 3026 QSOs, with BH masses ranging within 107<MBH< 1011M .
4.2.2
Mid-infrared and Far-infrared photometry
For our analysis we make use of the psf-smoothed and background subtracted PACS and SPIRE image products provided by the H-ATLAS team (Valiante et al. in prep) for the four fields of GAMA09 (54 deg2), GAMA12 (54 deg2), GAMA15 (54 deg2), and NGP (150 deg2) that overlap with the SDSS survey. Detailed information on the construction of the images is presented in Valiante et al. (in prep). The images used in our analysis have had the large scale background subtracted (i.e., the cirrus emission), and each pixel contains the best estimate of the flux density of a point source at that position, making them ideal for stacking analysis. In addition to the images there are also noise maps available that provide the sum of the instrumental and confusion noise at each position (details in Valiante et al. in prep).
To define the MIR properties of our sample we use the WISE all-sky survey (Wright
et al. 2010; catalogue available at: http://vizier.u-strasbg.fr/viz-bin/VizieR-3?-source=II/311/wise). Using a radius of 1 arcsec we match to the optical positions of our QSO sample described
in §4.2.1, with a spurious match fraction of ∼0.4%. We find that more than 90% of our sources have a WISE counterpart. For our analysis we use the W3 and W4 bands at 12µm and 22µm respectively.
4.2.3
Radio data and classification
To determine the radio luminosities of our QSO sample we use the FIRST radio cata- logue (Becker et al. 1995) that covers the full sky area observed by SDSS, to a sensitivity of 1mJy. To define our radio luminous QSO sub-sample we matched the SDSS QSO catalogue to the FIRST catalogue using a 2” radius, to minimise the number of spurious matches, with a resulting spurious match fraction of ∼2.5%. We calculate the 1.4GHz luminosity (L1.4GHz) from the catalogued flux densities, using the following equation:
L1.4GHz= 4πD2F1.4GHz(1 + z)−(1−α) (4.2.1)
where D is the luminosity distance, F1.4GHZ is the catalogued flux density, and assuming
fv∝ v−α with a spectral index of α =0.8. In Figure 2 we plot the radio luminosity of the detected sources as a function of redshift.
We classify sources as radio-luminous AGN, using a luminosity lower limit cut of L1.4GHz>1024W Hz−1 for z <0.8, and L1.4GHz >1025W Hz−1 for z >0.8 (see Fig. 2).
Based on work from McAlpine et al. (2013), Magliocchetti et al. (2016) argue that the radio luminosity beyond which the radio emission is dominated by the AGN evolves with redshift up to a redshift of z ∼ 1.8, after which it remains constant at L1.4GHz,limit=
1023.5W Hz−1. Our luminosity cut is always higher in comparison, meaning that we are selecting only AGN-powered radio sources, and that we are selecting the most powerful of radio AGN. Furthermore, in Figure 4.10 of section 4.4.3 we demonstrate how the radio luminosities of this sample are >1–3 orders of magnitude higher than the radio luminosi- ties predicted from the IR luminosities due to star-formation.
Within the redshift range studied here (z = 0.2–2.5), there are 258 QSOs classified as radio-luminous.